Modeling and Estimating the Effects of Institutional Variables on a Pay-as-you-go Social Security System and on Household Saving

Abstract

Using a two-period overlapping generations model and three panel data sets of annual aggregate data from twenty-five countries, we estimate a fixed-effects Euler equation for household saving. We focus on the effects of several institutional and other variables, such as corruption and the debt to gross domestic product (GDP) ratio, on household saving and on the probability that a pay-as-you-go social security system will grant pensions. We find that social security contributions reduce saving in a less than one-for-one manner. Also, as corruption or the debt to GDP ratio increases, the probability that the system will grant pensions falls, and so does the effect of social security contributions on saving. Finally, the marginal effect of an improvement in the quality of institutions on the credibility of the social security system is greater in countries where the quality of institutions is low than in countries where it is high, a result that stresses the role of institutions in reducing uncertainty about pensions.